With wider investigations and applications of autonomous robotics intelligent vehicles, the inverted pendulum on a cart (IPC) system has become more attractive for numerous researchers due to its concise representative structure. In this article, tracking-control problem IPC is considered investigated. Based Zhang dynamics (ZD) gradient (GD), novel kind ZG controllers are developed investigated achieving purpose, which contains z2g0 z2g1 types according number times using ZD GD methods....

Nonuniform thermal behavior in lithium-ion battery packs can accelerate aging, leading to inconsistent cell performance. If not adequately monitored and managed, this heating give rise unwanted side reactions, fires, explosions, underscoring the criticality of temperature field reconstruction. In recent years, data-driven methods have gained popularity for addressing reconstruction problem. However, many existing approaches require retraining when system parameters change, such as initial...

A new type of feed-forward 2-input neuronet using Chebyshev polynomials Class 1 (2INCP1) is constructed and investigated in this paper. In addition, with the weights-direct-determination method exploited to obtain optimal weights from hidden layer output directly (i.e., just one step), a structure-automatic-determination called weights-and-structure-determination (WASD) algorithm proposed determine number hidden-layer neurons 2INCP1. Such WASD includes procedure pruning (after net grows up)....

Zhang dynamics (ZD), which is based on an indefinite error-monitoring function called Zhangian, a powerful type of for online time-varying problems solving. Besides, gradient (GD), was originally designed to solve constant problems, can be generalized In this paper, the tracking-control problem general nonlinear system with fully trackable and partially paths presented investigated. Then, by combining ZD GD, innovative method ZD-GD proposed problem. Simulation results further substantiate...

Inverse kinematics (or say, redundancy resolution) is a fundamental issue in operating redundant manipulators, which has been widely investigated the past three decades. Many redundancy-resolution schemes have proposed for such problem solving, mainly work at single level (e.g., joint-velocity or joint-acceleration level). In this paper, robotic resolution, two general scheme-formulations different levels are presented and investigated, one of corresponds to velocity-level other...

Based on the theory of polynomial-interpolation and curve-fitting, a new multiple-input feed-forward neural network activated by Hermit orthogonal polynomials is proposed investigated. Besides, design makes (MIHOPNN) have no weakness dimension explosion. To determine optimal weights MIHOPNN, weight direct determination (WDD) method presented. obtain structure so-called (WASD) finally proposed, which aims at achieving best approximation accuracy while obtaining minimal number hidden-layer...

The pseudoinverse-based method, as a conventional is generally used in the motion planning and control of velocity and/or acceleration level(s) to solve complicated inverse-kinematics (IK) problem. In this paper, we firstly propose simple solution at jerk level IK problem by using direct derivative dynamics (DDD or D <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> for short) thrice without Zhang (ZD), which thus termed D3Z0 type. Note...

As an important issue of inverse kinematics, the self-motion planning and control (SMPC) redundant manipulators has been widely investigated to perform various subtasks. But there are still some weaknesses in SMPC researches [e.g., physical limits usually not taken into consideration, joint-velocity profiles mutational (or say rigorously, nondifferentiable) at time instants], which may jeopardize its future practical applications. Aiming remedying these mentioned above, a smoother (S-SMPC)...

Different from the Zhang function (i.e., a type of error function) defined in authors' previous work, new is proposed, designed and exploited to construct neural network (ZNN) models this paper. Moreover, two discrete-time ZNN are developed investigated solve problem time-varying linear equations (TVLE). Such can exploit methodologically time derivatives coefficients theoretical inverse coefficient matrix. When matrix positive-definite symmetric, BFGS quasi-Newton method introduced eliminate...

For returning to a near initial location for mobile redundant manipulator after task execution, wheel-joint coordinated minimum movement (WJCMM) scheme is proposed minimize the movement/displacement between final state and of robot when end-effector fulfills given task. Such can both coordinate simultaneously wheels platform joints complete task, achieve minimization. Besides, physical limits be incorporated into WJCMM scheme. Then, reformulated as quadratic program (QP) subject equality...

Zhang neural network (ZNN) is a novel class of recurrent with superior solution ability and convergence performance. For real-time Moore-Penrose pseudoinverses time-varying matrices based on continuous-time network, this paper proposes three different ZNN models, each which derived from specifically-chosen function (ZF). Theoretical analyses guarantee the global models their fast rate. Besides, proposed show additional great advantages when used to deal contrasting numbers rows columns....

Static reciprocal solving has been studied for many years. In this paper, we focus on time-varying (termed Zhang reciprocal) solving, i.e., regard the problem as a process. As both dynamics and gradient have proven powerful tools problems propose novel continuous-time Zhang-gradient (ZG) type (ZR) model solving. Through computer simulations, further verify effectiveness of ZG-type ZR model.

Large language models (LLMs) commonly employ autoregressive generation during inference, leading to high memory bandwidth demand and consequently extended latency. To mitigate this inefficiency, we present Bi-directional Tuning for lossless Acceleration (BiTA), an innovative method expediting LLMs via streamlined semi-autoregressive draft verification. Inspired by the concept of prompt tuning, enhance with a parameter-efficient design called bi-directional tuning capability in generation....

A novel class of neural dynamics, Zhang dynamics (ZD), has been recently proposed for online time-varying problem solving. The design method ZD is based on an indefinite error-monitoring function called the (ZF), instead conventional norm-based scalar-valued energy function. In this paper, different models ZFs are and developed solving inverse square root (TVISR) problem. addition, paper investigates modeling using MATLAB Simulink techniques. Results substantiate efficacy superiority TVISR

The forward kinematics for robotic manipulators has a mathematically analytical form, while the inverse-kinematics problem is complicated. There are many methods to solve problem; such as conventional pseudo-inverse based method, which general but some problems and limitations. Therefore, it worth investigating solution of using dynamic methods. Based on previous researches, consist Zhang dynamics (ZD) gradient (GD). In this paper, by ZD twice not GD, controller named Z2G0 investigated...